Results indicate that the LSI-R:SV is a useful tool for assisting in the prediction of both nonviolent and violent recidivism by mentally ill offenders. Beyond predicating recidivism, it has utility in identifying the domains of dynamic variables that need to be addressed to assist in the rehabilitation of mentally ill offenders. Further analysis more specifically showed that as expected, the LSI-R:SV was particularly useful in aiding in the prediction of both nonviolent and violent recidivism in a subpopulation of mentally ill offenders who did not abuse alcohol and illicit substances. However, the LSI-R:SV was shown to be a poor tool for the assessment of risk of recidivism in a subpopulation of dually diagnosed offenders who abused alcohol and illicit drugs at a level that was severe enough to interfere with the undertaking of their occupation (work or school). Data were collected from 208 individuals admitted to Thomas Embling Hospital, Victoria’s secure forensic mental health facility operated by the Victorian Institute of Forensic Mental Health, and with a discharge date prior to November 2003. Tables, figure, and references